f87689a4ef
Root cause from latest docker build log: ValueError: You set image=0 in --limit-mm-per-prompt, but found 1 items → Engine background task crashes → AsyncEngineDeadError → all subsequent 503 Fixes: 1. computility-run.yaml: add --limit-mm-per-prompt image=1 Prevents multimodal ValueError from killing the engine process. 2. patch_ops.sh: DON'T overwrite base image's corex_gdn.py/corex_moe.py Comp 168 log proves base image's corex modules work with libcorex_gdn.so. Our overwrite broke CoreXGDN.__init__ (unexpected kwarg 'num_v_heads'). Only deploy ours if base has NO corex modules at all. Also deploy corex_fa2.py if base lacks it. 3. qwen3_5.py: try multiple CoreXGDN init signatures Base image CoreXGDN may accept different kwargs than ours. Try kwargs form first, fall back to positional. 4. corex_gdn.py: accept both calling conventions in __init__ Future-proof for when we DO need to deploy ours. 5. Copied upstream_ref headers: ilu_layer_fused_moe.h, ilu_layer_attention.h Last 2 missing ILU files from xllm. All 14/14 now present.
83 lines
2.6 KiB
C++
83 lines
2.6 KiB
C++
/* Copyright 2025 The xLLM Authors. All Rights Reserved.
|
|
|
|
Licensed under the Apache License, Version 2.0 (the "License");
|
|
you may not use this file except in compliance with the License.
|
|
You may obtain a copy of the License at
|
|
|
|
https://github.com/jd-opensource/xllm/blob/main/LICENSE
|
|
|
|
Unless required by applicable law or agreed to in writing, software
|
|
distributed under the License is distributed on an "AS IS" BASIS,
|
|
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
|
See the License for the specific language governing permissions and
|
|
limitations under the License.
|
|
==============================================================================*/
|
|
|
|
#pragma once
|
|
|
|
#include <torch/torch.h>
|
|
|
|
#include <tuple>
|
|
|
|
#include "framework/kv_cache/kv_cache.h"
|
|
#include "framework/model/model_input_params.h"
|
|
#include "layers/common/attention_metadata.h"
|
|
|
|
namespace xllm {
|
|
namespace layer {
|
|
class AttentionImpl : public torch::nn::Module {
|
|
public:
|
|
AttentionImpl() = default;
|
|
|
|
AttentionImpl(int64_t num_heads,
|
|
int64_t head_size,
|
|
float scale,
|
|
int64_t num_kv_heads,
|
|
int64_t sliding_window);
|
|
AttentionImpl(int64_t num_heads,
|
|
int64_t head_size,
|
|
int64_t num_kv_heads,
|
|
int64_t v_head_dim,
|
|
int64_t sliding_window,
|
|
float scale,
|
|
bool use_fused_mla_qkv,
|
|
bool enable_lighting_indexer,
|
|
bool enable_mla);
|
|
|
|
std::tuple<torch::Tensor, std::optional<torch::Tensor>> forward(
|
|
const AttentionMetadata& attn_metadata,
|
|
torch::Tensor& query,
|
|
torch::Tensor& key,
|
|
torch::Tensor& value,
|
|
KVCache& kv_cache);
|
|
|
|
void prefill_forward(torch::Tensor& query,
|
|
torch::Tensor& key,
|
|
torch::Tensor& value,
|
|
torch::Tensor& output,
|
|
const torch::Tensor& k_cache,
|
|
const std::optional<torch::Tensor>& v_cache,
|
|
const AttentionMetadata& attn_metadata);
|
|
|
|
void decoder_forward(torch::Tensor& query,
|
|
torch::Tensor& output,
|
|
const torch::Tensor& k_cache,
|
|
const std::optional<torch::Tensor>& v_cache,
|
|
const AttentionMetadata& attn_metadata);
|
|
|
|
private:
|
|
int64_t num_heads_;
|
|
int64_t head_size_;
|
|
float scale_;
|
|
int64_t num_kv_heads_;
|
|
int64_t v_head_dim_;
|
|
bool use_fused_mla_qkv_;
|
|
bool enable_lighting_indexer_;
|
|
bool enable_mla_;
|
|
int64_t sliding_window_;
|
|
};
|
|
TORCH_MODULE(Attention);
|
|
|
|
} // namespace layer
|
|
} // namespace xllm
|